A Practice Has No Users: Arguments Against Instrumental Living
Note: Some articles belong to one site. These belong to all of them. Superconnectors are pieces written at the intersections. Where opera meets product thinking, where walking meets AI, where the archive meets the self.
The autonomousmatt.com conversation product has a system prompt. It enforces an essayistic first-person voice. It prohibits bullet points and em-dashes. It routes queries through a RAG pipeline, reformulates them with a smaller model, retrieves semantically proximate content from an Upstash vector store, and generates a response that sounds like the person who built it. The whole system is designed for users who arrive without context, who need onboarding, whose sessions are logged in Redis, whose queries reveal what the corpus is missing.
Aufbruchmatt has none of this. There is no onboarding. There is no retention metric. There is no feature roadmap driven by user research. There is one person who walks, and the same person who writes about it, and the same person who reads what was written, and they are all the same person at different points in the same year. This should be a trivial distinction. It is not.
Product thinking is, at its foundation, a theory of the gap. There is a person with a need, and a system that addresses it, and the product lives in the space between them. Every tool of the discipline, user stories, north star metrics, feedback loops, A/B tests, exists to manage and eventually close that gap. The product succeeds when the user's need is met. It fails when the gap remains. The whole apparatus assumes that the maker and the user are different people, and that the distance between them is the problem to be solved. When the maker and the user are the same person, the gap collapses. And something strange happens in the collapse.
Schopenhauer argued that the will is never finally satisfied by its object. You want a thing. You get it. The wanting relocates to a new object. This is not a failure of the will. It is the will's nature. Desire is not a problem with a solution. It is a structure. The product manager who believes they can achieve product-market fit so perfect that users stop wanting new features has not understood Schopenhauer. There is always another feature. There is always another unmet need. The gap cannot be closed because the gap is constitutive of the relationship between maker and user. The scored walk does not have this problem. Or rather, it has a different version of it.
Each month begins with uncertainty. Which route, which passages, which quality of attention. Each walk ends with something that was not anticipated at the start. Not a product improvement. Not a resolved user need. Something closer to what the Zen tradition calls shoshin, beginner's mind. The walk does not accumulate toward mastery in the way a product accumulates toward maturity. It begins again. Holländer in January. Meistersinger in May. The calendar turns and the practice resets, carrying only what the body has learned, which is not the same as what the mind has recorded.
This is what a practice is. A practice is a form of activity whose value is not separable from its continuation. You do not practice the piano in order to have practiced it. You practice in order to be, at this moment, practicing. The product that achieves perfection can be deprecated. The practice that achieves perfection has simply deepened. Here is the uncomfortable implication. The autonomousmatt.com products, the conversation interface, the visual inspiration tool, the Ariadne-like cross-constellation search, are built on the assumption that discovery is the value. The user arrives, queries, finds. The Voyage AI embeddings do their work. The vector proximity scores surface what is relevant. The system is optimized for the moment of arrival. The user who does not yet know what is there.
But the person who built those products already knows what is there. He wrote it. He ingested it. He has been walking through it, literally, for months. The RAG pipeline retrieves, for him, a reflection of his own thought. Arranged differently, surfaced from different angles, but fundamentally familiar. He is not discovering the corpus. He is being shown himself. This is not a failure of the product. It is a revelation about what the product actually does, for this user, in this case.
The autonomous/matt conversation system is, for its builder, a practice tool, not a discovery tool. It does not tell him what he does not know. It tells him what he knows, slowly, in an order he did not choose. Which is structurally identical to what the scored walk does. The Furtwängler Tristan does not tell you things about longing you did not already feel. It tells you what you already felt, arranged in an order, Act I, Act II, Act III, that you did not determine, at a pace, two hours, three hours, the long suspension of the second act, that is not yours to control.
The product and the practice converge, in this account, on the same activity. The deliberate encounter with your own interiority, mediated by a structure you built but cannot fully control. The difference is that the product imagines an other. The practice knows it is alone.
There is a specific moment on the May Meistersinger walks. Somewhere in the third act, when the Preislied arrives and the whole opera's argument about form and feeling resolves. When the question of audience becomes irrelevant. Not because no one is watching. Because the question itself has dissolved. The walk is happening. The music is happening. The city is happening. There is no gap to close. There is no user to satisfy. That is what a practice feels like from the inside. It does not feel like the absence of a product. It feels like the product was always pointing toward this.
